Vol. 10: 79–88, 2018 AQUACULTURE ENVIRONMENT INTERACTIONS Published February 19 https://doi.org/10.3354/aei00255 Aquacult Environ Interact

OPENPEN FEATURE ARTICLE ACCESSCCESS

Adding value to ragworms (Hediste diversicolor) through the bioremediation of a super-intensive marine fish farm

Bruna Marques1, Ana Isabel Lillebø1, Fernando Ricardo1, Cláudia Nunes2,3, Manuel A. Coimbra3, Ricardo Calado1,*

1Department of Biology & CESAM & ECOMARE, 2CICECO − Aveiro Institute of Materials, and 3Department of Chemistry & QOPNA, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal

ABSTRACT: The aim of this study was to evaluate the potential added value of Hediste diversicolor, cultured for 5 mo in sand bed tanks supplied with effluent wa- ter from a super-intensive marine fish farm, by com- paring their fatty acid (FA) profile with that of wild specimens. The polychaetes showed an approximately 35-fold increase in biomass during the experimental period and their FA profile was significantly different from that of wild specimens. In cultivated specimens, the most abundant FA class was that of highly unsatu- rated FA (HUFA), with eicosapentaenoic acid (EPA, 20:5n-3) being the best represented. Similar percent- age (SIMPER) analysis showed an average 20.2% dis- similarity between the FA profile of wild and culti- vated specimens, supporting the view that the culture system employed enables the recovery of high value Ragworms (Hediste diversicolor) on sand filters of a super- nutrients (e.g. EPA and docosahexaenoic acid [DHA, intensive brackish-water fish farm, cultured using the 22:6n-3]) from fish feeds into the tissues of H. diversi- farm’s organic-rich effluent and displaying a greater con- color that would otherwise be lost from the production tent of docosahexaenoic acid (DHA, 22:6n-3) than wild environment. While the nutritional value of wild rag- conspecifics. worms is well established in marine aquaculture Photo: Bruna Marques, Universidade de Aveiro (namely for broodstock maturation diets), the higher level of DHA displayed by the specimens produced under the proposed culture system may grant them a premium market value. where’ (Martins et al. 2010). The principles behind RAS promote the treatment and reuse of culture wa- KEY WORDS: Integrated multi-trophic aquaculture · ter, with 10% or less of the total water volumes hav- IMTA · Polychaete-assisted sand filters · Fatty acids ing to be replaced per day (Hutchinson et al. 2004, van Rijn et al. 2006). However, one of the constraints impairing the expansion of these production systems (in a closed or semi-closed operation) is the challenge INTRODUCTION associated with the load of organic-rich suspended particulate matter (POM), dissolved orga nic matter Recirculating aquaculture systems (RAS) are cur- (DOM) and nutrients in dissolved inorganic form (ni- rently considered one of the paradigms of the Blue trogen, N, and phosphorus, P) present in its effluent Revolution, as they allow people to ‘grow fish any- (Schneider et al. 2005, Marques et al. 2017).

© The authors 2018. Open Access under Creative Commons by *Corresponding author: [email protected] Attribution Licence. Use, distribution and reproduction are un - restricted. Authors and original publication must be credited. Publisher: Inter-Research · www.int-res.com 80 Aquacult Environ Interact 10: 79–88, 2018

A number of integrated multi-trophic aquaculture and they are also able to scavenge and actively prey (IMTA) systems have been developed in order to re - on other organisms (including conspecifics) (Luis & duce the nutrient load present in the effluents pro- Passos 1995, Fidalgo e Costa et al. 2000, Bischoff et duced when growing finfish or shrimp. In other al. 2009). The interactions with its environment show words, IMTA combines the integrated culture of fed an efficient adaptation to a variation of environmen- target species (e.g. finfish/shrimp) with that of ex- tal parameters such as salinity, temperature and dis- trac tive species that use particulate or dissolved solved oxygen (Smith 1964, Fritzsche & von Oertzen orga nic matter (e.g. shellfish/herbivorous fish) or dis- 1995, Murray et al. 2017). Polychaetes also play an solved inorganic nutrients (e.g. seaweed/halophytes) important ecological role in the marine environment, generated through the excretion products of fed spe- as they are major contributors to the resuspension of cies and uneaten feed (Schneider et al. 2005, Chopin organic matter via bioturbation (Würzberg et al. et al. 2008, Alexander et al. 2015). Overall, this envi- 2011), a feature of great relevance for their potential ronmentally friendly approach aims to address the use in IMTA. Indeed, some studies have already im pacts commonly associated with conventional highlighted how polychaetes can be successfully aqua culture, such as nutrient loading and sedimen- employed in the bioremediation of aquaculture efflu- tation (Schneider et al. 2005, Barrington et al. 2009). ents under an IMTA framework (Palmer 2010, Fang In this way, promoting IMTA practices may help to et al. 2017). overcome the current bottlenecks faced by enter- The nutritional value of ragworms is well estab- prises operating RAS, as the nutrient-rich effluents lished in marine aquaculture, with these organisms that they generate may be used to culture additional being a highly valued item in marine finfish and species. While on one side this approach allows the shrimp maturation diets (Olive 1999, Techaprem- nutrient load issue to be addressed from a biomitiga- preecha et al. 2011, Santos et al. 2016). One of the tion perspective, through the incorporation of nutri- main reasons for their popularity is their fatty acid ents in extractive species biomass, it also opens up (FA) profile, namely the levels they display of impor- the opportunity of adding new cash-crops to the pro- tant polyunsaturated FA (PUFA) (Brown et al. 2011, duction model through the valorization of those Santos et al. 2016). Some polychaete species, includ- extractive species (Chopin et al. 2008, Alexander et ing H. diversicolor, are able to biosynthesize PUFA, al. 2015). such as 20:5n-3 eicosapentaenoic acid (EPA) and The polychaete Hediste diversicolor (O.F. Müller, 22:6n-3 docosahexaenoic acid (DHA), which are 1776), popularly known as ragworm, is a candidate known to be essential for marine finfish and shrimp species for land-based IMTA systems, as it can effi- (Olive 1999, Fidalgo e Costa et al. 2000). ciently recycle particulate organic nutrients present In order to evaluate the potential valorization of in fish farm effluents (Scaps 2002, Bischoff et al. 2009, H. diversicolor cultured in sand bed tanks supplied Santos et al. 2016). This polychaete is a burrowing with effluent water from a super-intensive marine species that inhabits the soft bottoms of shallow mar- fish farm, their fatty acid profile was determined and ine and brackish waters environments, generally in compared with that of wild specimens. The experi- sediments with high organic contents, in the North mental procedure aimed to test (1) whether ragworm temperate zone of European and North American FA profiles depend on food source, by comparing Atlantic coasts (Scaps 2002, Lillebø et al. 2012). H. cultured and wild specimens, and (2) whether the FA diversicolor exhibits a ‘bentho-pelagic life cycle’ profiles of ragworms cultured in sand bed tanks sup- (Scaps 2002, Nesto et al. 2012). It is one of the rare plied with RAS effluent are size-dependent, by com- Nereid species which remain atokous throughout paring small, medium and large ragworms. their lifespan, contrasting with some other Nereids which undergo a metamorphosis to a typical epitok- ous heteronereid form (Scaps 2002, Breton et al. MATERIALS AND METHODS 2003, Aberson et al. 2011). During reproduction, the female incubates the eggs for a period of 10 to 14 d Experimental set-up inside the gallery, and dies soon afterwards. In this way, the direct or brooded larval development per- Effluent originating from a super intensive RAS mits a greater flexibility in the level of development system farming Senegalese sole Solea senegalensis of the released offspring (Aberson et al. 2011). The was pumped from a settling basin to a bio-block feeding modes of this polychaete are diversified, tower system, in order to allow the effluent water to ranging from surface deposit to suspension feeding, trickle and increase its oxygen levels before reaching Marques et al.: Ragworm cultivation and added value 81

the header tank reservoir. In this header tank, the (20). Subsequently, all sampled specimens were effluent was strongly aerated and set to flow in paral- separated into 3 pre-established size classes (small lel into 6 experimental sand bed tanks. Each of these <30 mm, medium 30 to 50 mm and large >50 mm). At tanks had an approximate volume of 1 m3 and a sur- the beginning of the experiment, triplicate samples of face area of 1 m2. The bottom of each tank was cov- WP (all from the large class) that were used to stock ered by 200 mm of sand (1 to 2 mm grain size substra- the tanks, were left to depurate overnight under tum) beneath which a draining pipe allowed the identical conditions as those previously described. effluent water to percolate through the sand bed. After depuration, wild polychaetes and those origi- Each tank was equipped with 2 outlets, one that reg- nating from the sand bed tanks were freeze-dried ulated the water level inside the tank and allowed and stored at −80°C for subsequent FA analysis. the water to drain and another one that was set to prevent the tank from overflowing in case the sand bed became clogged and impaired water percola- Sampling of potential nutrient sources of POM for tion. Sand bed tanks were placed in parallel and sup- FA analysis plied with aerated RAS effluent water by gravity at a flow of 180 l h−1. All tanks were equipped with a To perform the FA characterization of all potential 0.8 m diameter ring of aeration hose supplying com- nutrient sources available to the ragworms stocked in pressed air (see Marques et al. 2017). Three of the 6 the sand bed tanks, triplicate samples of the 2 fish experimental sand bed tanks were stocked with wild feed types (FEED A and B) used during the grow-out polychaetes (WP), while the remaining 3 were set as of S. senegalensis in the RAS system, as well as controls. The experiment was run for 5 mo (from May uneaten/undigested feed and fish feces accumulated to October) and no additional feed was provided to in the cyclone filters (POM), were collected, freeze the polychaetes during this period to supplement the dried and stored at −80°C for posterior FA analysis. organic-rich effluent containing uneaten/undigested The top 10 mm of the sand bed surface of each tank feed and fish feces. During the experimental period, stocked with ragworms, all organic-rich settled parti- temperature, salinity, pH and dissolved oxygen con- cles (OM), were also collected in triplicate from each centration of the effluent water being supplied to the tank and were also processed as described above for tanks was monitored twice a month using a WTW FA analysis. Cond 3110/SET 1 equipped with TetraCon® 325, a WTW pH 330i/SET equipped with SenTix® 41, and a WTW Oxi 3210/SET 2 equipped with CellOx® 325-3, FA extraction and analysis respectively. The derivation of FAs for gas chromatography (GC) analysis was performed following the methodology Polychaete stocking and sampling described by Aued-Pimentel et al. (2004), adjusting the weight, as this method has the advantage of being Wild specimens of Hediste diversicolor (average performed at room temperature and thus reducing total length between 80 and 100 mm, the commercial the risks of FA decomposition. All freeze-dried sam- size of this species when traded as bait for sports fish- ples were powdered and homogenized, weighed ac- ing) were collected in the Ria de Aveiro coastal curately in a Sovirel Pyrex glass tube (40 mg of rag- lagoon (Portugal 40° 38’ 04.8” N, 8° 39’ 52.4” W) by worm biomass, 20 mg of FEED, ~40 mg of POM and local fishermen, and 200 individuals (with a com- 150 mg of OM) and dissolved in 1 ml of the internal bined weight of 130 g) were stocked in each of the 3 standard solution of a methyl ester fatty acid C21:0 in sand bed tanks randomly selected to act as poly- n-hexane (0.35 g l−1). In the same tube, 0.2 ml of a chaete-assisted sand filters. At 5 mo post-stocking, methalonic KOH solution (2 mol l−1) was added, and polychaetes were sampled using a hand corer the tube was sealed and mixed vigorously in a vortex (110 mm diameter and 150 mm long), with 3 cores shaker for 30 s. Following this procedure, 2 ml of a sat- being taken from each colonized sand bed tank. urated NaCl solution was added to the tube, and the Polychaete specimens were sorted in situ and trans- mixture was centrifuged for 5 min at 1409 × g. The ported to the laboratory in sterilized sand and clean separated organic phase (1 ml) was transferred into seawater. At the laboratory, all specimens were left another tube and the excess solvent was removed to depurate overnight in pre-combusted sand and under vacuum. The oil obtained was dissolved in n- artificial seawater prepared to match salinity in situ hexane (200 µl) and analyzed by gas chromatography 82 Aquacult Environ Interact 10: 79–88, 2018

with a flame ionization detector (GC-FID), using a detailed description of all the statistical analysis Perkin Elmer 400 instrument (PerkinElmer). The de- described above please refer to Clarke & Gorley tector and injector were kept at 250°C, with hydrogen (2006). as carrier gas. FAs were separated in a fused-silica capillary column, DB-FFAP (30 m length, 0.32 mm in- ternal diameter, 0.25 µm film thickness, J & W Scien- RESULTS tifics) with the following temperature program: 50°C for 3 min, 40°C min−1 to 160°C, 2°C min−1 to 210°C, Biomass production of ragworms 20°C min−1 to 250°C (for 1 min). The identification of FAs was done by matching with a previously injected During the 5 mo experimental period, water para - standards mixture (Supelco® 37 component FAME meters (average ± standard deviation, SD) within the mix, Sigma-Aldrich). The FA content (µg mg−1 dry sand bed tanks remained stable (see Table A1 in the weight, DW) in the samples analyzed was calculated Appendix), with an average water temperature of considering the relationship between mass, the area 19.6 ± 1.3°C, salinity 21.2 ± 0.2, pH 7.8 ± 0.2 and dis- of FAs and the internal standard (C21:0). PUFA are solved oxygen 8.4 ± 0.7 mg l−1. defined as all FA with ≥2 double bonds; in the present The final average (±SD) weight of cultured poly- study, highly unsaturated fatty acids (HUFA, FAs with chaete biomass was 2622 ± 869 g, corresponding to ≥4 double bonds) are considered separately. 104 ± 68, 226 ± 137 and 2292 ± 664 g for SPC, MPC and LPC, respectively. During this period, polychaete density increased from 200 individuals (ind.) m−2 to Statistical analysis up to 7094 ind. m−2, which represented an approxi- mately 35-fold increase in density, solely considering Statistical analysis was performed using PRIMER LPC (initial biomass of 130 g increasing to 2292 g 5 v6 with the PERMANOVA+ add-on. A resemblance mo later) and approximately 18-fold considering the matrix using the content (µg g−1 DW) of each FA in whole biomass of cultured polychaetes (initial bio- each sample was prepared using the Bray-Curtis mass of 130 g increasing to 2622 g 5 mo later). similarity coefficient, after performing a log(x + 1) On average, the total numbers of polychaetes per transformation to emphasize compositional rather tank were 9608 ± 5922, 3374 ± 1464 and 7094 ± than quantitative differences (Anderson 2008). Hy - 1375 ind. m−2, for SPC, MPC and LPC, respectively. potheses were tested by performing 2 independent The average number of LPC recorded represented 1-way analyses of similarities (ANOSIM). To assess approximately 100 bait packages similar to those the differences between FA profiles of cultured originally introduced into the system (each pack of polychaetes versus WP and those of small, medium bait holds ~70 individuals). and large cultured polychaetes (SPC, MPC and LPC, respectively) in sand bed tanks, a global R sta- tistic was calculated where values close to 1 indicate FA profile analysis maximum differences between groups and values near 0 indicate a complete group overlay. Similarity The FA profiles of WP, SPC, MPC and LPC (these percentage (SIMPER) analysis was also performed displaying a similar size to that of wild specimens), to evaluate the percentage that each FA contributed POM, OM, and FEED A and B are summarized in to the dissimilarity recorded between samples, with Table 1. Considering FA classes, WP and LPC dis- those contributing 50% of cumulative dissimilarities played similar profiles; however, the ANOSIM test being highlighted. Hierarchical cluster analysis was revealed the existence of significant differences (R = performed to group the samples according to their 1, p = 0.002) between the FA profiles of WP and LPC. similarity. A dendrogram was used to highlight the HUFA was the most abundant FA class (8.78 ± 0.70 hierarchical similarity among the samples. Similarity and 14.23 ± 1.08 µg g−1 DW for WP and LPC, respec- among the samples was calculated using the Bray- tively), with EPA (20:5n-3) being the best represented Curtis similarity measure and the group average FA (5.51 ± 0.29 and 8.34 ± 0.36 µg g−1 DW for WP and algorithm was used to group the samples succes- LPC, respectively). PUFA averaged 2.34 ± 0.38 and sively in a hierarchical way. A canonical analysis of 3.99 ± 0.46 µg g−1 DW for WP and LPC, respectively, principal coordinates (CAP) was performed to eval- while monounsaturated fatty acids (MUFA) averaged uate the strongest correlation of LPC in the pre- 6.69 ± 1.28 and 10.05 ± 1.08 µg g−1 DW for WP and defined groups (OM, POM, FEED A and B). For a LPC, respectively. Concerning saturated fatty acids Marques et al.: Ragworm cultivation and added value 83

Table 1. Fatty acid (FA) profiles (µg g−1 DW) of wild (WP), and small, medium and large Hediste diversicolor cultured in the sand bed tanks (SPC, MPC and LPC, respectively), organic matter (OM), particulate organic matter (POM) and fish feed (FEED A and FEED B). Values are averages of 9, 6 or 3 replicates ± standard deviation. ND: fatty acid not detected. SFA: satu- rated fatty acids (14:0; 15:0; 16:0; 17:0; 18:0; 20:0); MUFA: monounsaturated fatty acids (16:1n-7; 16:1n-9; 18:1n-9; 18:1n-7; 18:1n-5; 20:1n-9; 22:1n-9); PUFA: polyunsaturated fatty acids (16:2n-6; 18:2n-6; 18:2n-3; 18:3n-6; 18:3n-3; 20:2n-6; 20:2n-9; 20:3n-6); HUFA: highly unsaturated fatty acids (20:4n-6 [arachidonic acid, AA]; 20:5n-3 [eicosapentaenoric acid, EPA]; 22:4n- 6; 22:5n-3; 22:6n-3 [docosahexaenoic acid, DHA]). PUFA are defined as all FA with ≥2 double bonds; in the present study, HUFA (FA with ≥4 double bonds) are not considered within ∑PUFA. Others refers to the FAs 14:0 (iso), 15:0 (iso) and 16:0 (iso)

FA WP SPC MPC LPC OM POM FEED A FEED B (n = 9) (n = 9) (n = 9) (n = 9) (n = 9) (n = 6) (n = 3) (n = 3)

14:0 0.10 ± 0.03 0.22 ± 0.03 0.25 ± 0.04 0.33 ± 0.05 0.36 ± 0.03 2.39 ± 0.37 4.09 ± 0.44 8.18 ± 0.63 15:0 0.27 ± 0.05 0.23 ± 0.03 0.21 ± 0.03 0.28 ± 0.03 0.06 ± 0.01 0.22 ± 0.03 ND ND 16:0 4.22 ± 1.10 3.53 ± 0.28 3.66 ± 0.57 5.63 ± 0.85 1.31 ± 0.22 8.16 ± 1.16 13.53 ± 1.33 23.44 ± 1.22 17:0 0.27 ± 0.06 0.34 ± 0.03 0.34 ± 0.02 0.46 ± 0.07 ND 0.35 ± 0.09 0.60 ± 0.06 1.48 ± 0.21 18:0 1.65 ± 0.37 1.56 ± 0.19 1.64 ± 0.13 2.28 ± 0.16 0.34 ± 0.06 1.87 ± 0.23 0.61 ± 0.03 3.39 ± 1.21 ∑SFA 6.52 ± 1.62 5.89 ± 0.55 6.10 ± 0.79 8.99 ± 1.16 2.14 ± 0.36 13.16 ± 1.89 18.83 ± 1.86 36.50 ± 3.27 16:1n-7 0.27 ± 0.06 0.57 ± 0.08 0.63 ± 0.07 0.90 ± 0.13 0.14 ± 0.20 2.38 ± 0.40 ND ND 16:1n-9 ND ND ND ND 0.86 ± 0.16 ND 4.66 ± 0.44 9.09 ± 0.76 18:1n-9 1.96 ± 0.43 1.83 ± 0.22 1.96 ± 0.06 2.39 ± 0.19 0.48 ± 0.08 7.77 ± 1.95 2.97 ± 0.31 9.64 ± 5.07 18:1n-7 0.31 ± 0.05 0.77 ± 0.06 0.89 ± 0.20 1.39 ± 0.20 0.58 ± 0.10 1.53 ± 0.63 25.13 ± 3.17 5.98 ± 4.13 18:1n-5 1.74 ± 0.44 1.49 ± 0.20 1.43 ± 0.11 1.87 ± 0.22 ND ND ND ND 20:1n-9 1.86 ± 0.22 2.14 ± 0.24 2.27 ± 0.17 2.80 ± 0.20 0.14 ± 0.18 0.86 ± 0.17 2.71 ± 0.25 1.99 ± 0.10 22:1n-9 0.54 ± 0.08 0.48 ± 0.09 1.00 ± 0.12 0.71 ± 0.11 0.05 ± 0.02 1.64 ± 0.60 2.00 ± 0.21 2.11 ± 0.19 ∑MUFA 6.69 ± 1.28 7.28 ± 0.90 8.18 ± 0.73 10.05 ± 1.08 2.26 ± 0.74 14.17 ± 3.74 37.47 ± 4.38 28.81 ± 10.25 18:2n-6 0.22 ± 0.03 0.56 ± 0.08 0.72 ± 0.16 1.13 ± 0.09 0.16 ± 0.03 4.31 ± 1.10 9.68 ± 1.13 5.31 ± 0.40 18:3n-6 1.12 ± 0.15 0.14 ± 0.05 0.17 ± 0.01 0.32 ± 0.04 0.09 ± 0.02 0.65 ± 0.20 2.75 ± 0.26 1.62 ± 0.12 18:3n-3 0.49 ± 0.10 ND ND ND ND 0.29 ± 0.10 ND ND 20:2n-6 0.25 ± 0.05 0.49 ± 0.05 0.56 ± 0.09 1.08 ± 0.24 2.29 ± 0.20 ND 0.18 ± 0.31 0.11 ± 0.08 ∑PUFA 2.34 ± 0.38 2.08 ± 0.33 1.89 ± 0.33 3.99 ± 0.46 2.55 ± 0.25 5.26 ± 1.40 14.53 ± 1.86 10.79 ± 1.16 20:4n-6 (AA) 0.84 ± 0.10 0.99 ± 0.33 1.25 ± 0.15 1.61 ± 0.17 0.12 ± 0.02 0.15 ± 0.04 0.74 ± 0.13 1.10 ± 0.04 20:5n-3 (EPA) 5.51 ± 0.29 5.65 ± 0.62 6.43 ± 0.51 8.34 ± 0.36 0.36 ± 0.12 0.11 ± 0.04 7.06 ± 0.71 16.20 ± 2.15 22:4n-6 0.39 ± 0.19 1.13 ± 0.26 1.24 ± 0.04 1.75 ± 0.23 ND 0.31 ± 0.07 0.17 ± 0.02 0.31 ± 0.02 22:5n-3 2.03 ± 0.12 1.25 ± 0.16 1.37 ± 0.17 1.72 ± 0.18 0.04 ± 0.01 0.22 ± 0.09 1.25 ± 0.16 1.32 ± 0.21 22:6n-3 (DHA) ND 0.60 ± 0.19 0.49 ± 0.07 0.82 ± 0.14 0.37 ± 0.06 1.73 ± 0.58 8.25 ± 0.75 10.61 ± 1.29 ∑HUFA 8.78 ± 0.70 9.62 ± 1.56 10.79 ± 0.92 14.23 ± 1.08 0.89 ± 0.21 2.52 ± 0.83 17.48 ± 1.77 29.54 ± 3.70 ∑Others 0.09 ± 0.00 0.51 ± 0.10 0.43 ± 0.08 0.38 ± 0.08 0.21 ± 0.11 0.22 ± 0.02 ND ND

(SFA), WP presented 6.52 ± 1.62 µg g−1 DW, while LPC explained by the following FAs: 18:3n-3 (alpha- displayed 8.99 ± 1.16 µg g−1 DW, with the most repre- linolenic acid, ALA), 22:4n-6 (docosatetraen oic acid, sentative SFA being palmitic acid (16:0) (4.22 ± 1.10 DTA), 18:2n-6 (linoleic acid), 22:6n-3 (DHA), 18:1n-9 and 5.63 ± 0.85 µg g−1 DW for WP and LPC, respec- (oleic acid), 18:3n-6 (gamma linolenic acid, GLA) and tively). The identification and quantification of the FA 22:1n-9 (erucic acid) (Table 2). profiles of FEED A and B, POM and OM revealed that The ANOSIM test revealed the existence of signif- the most representative FA in each of these samples icant differences between the FA profiles of SPC and was: (MUFA) vaccenic acid (18:1n-7) (25.13 ± 3.17 µg LPC (R = 0.474, p = 0.011), and MPC and LPC (R = g−1 DW) for FEED A; (SFA) palmitic acid (16:0) for 0.319, p = 0.017). However, the low R-values dis- FEED B (23.44 ± 1.22 µg g−1 DW) and POM (8.16 ± played in both comparisons suggest that these differ- 1.16 µg g−1 DW), and (PUFA) eicosadi enoic acid ences were more likely due to natural variability than (C20:2n-6) (2.29 ± 0.20 µg g−1 DW) for OM. promoted by the different size of the polychaetes. No SIMPER analysis showed that the FA profiles of WP significant differences were recorded in the FA pro- and LPC displayed an average dissimilarity of 20.2%, files of SPC vs. MPC (R = 0.131, p = 0.121). SIMPER with more than 50% of that dissimilarity being analysis (Table 2) revealed that the average dissimi- 84 Aquacult Environ Interact 10: 79–88, 2018

Table 2. SIMPER overall average dissimilarities (%) between the mean fatty acid (FA) profiles of wild (WP), and small, medium and large (SPC, MPC and LPC, respectively) Hediste diversicolor cultured in the sand bed tanks, and organic matter (OM), particulate organic matter (POM) and fish feed (FEED A and FEED B)

W & LPC SPC & LPC MPC & LPC SPC & MPC FA Contrib. Cum. FA Contrib. Cum. FA Contrib. Cum. FA Contrib. Cum. % % % % % % % %

18:3n-3 8.98 8.98 18:2n-6 9.53 9.53 16:0 8.87 8.87 20:4n-6 10.76 10.76 22:4n-6 8.54 17.53 22:5n-3 8.52 18.05 18:2n-6 8.24 17.11 20:5n-3 8.22 18.98 18:2n-6 8.30 25.83 20:4n-6 7.73 25.78 20:5n-3 7.67 24.79 16:0 7.31 26.29 22:6n-3 8.08 33.91 22:1n-9 7.41 33.19 22:1n-9 6.79 31.57 22:4n-6 6.93 33.23 18:1n-9 7.70 41.61 16:0 7.30 40.49 20:2n-6 6.76 38.34 22:6n-3 6.78 40.01 18:3n-6 6.85 48.46 20:2n-6 6.58 47.07 18:1n-9 6.63 44.97 18:2n-6 6.61 46.62 22:1n-9 6.51 54.97 18:1n-9 6.18 53.25 18:0 5.68 50.65 18:1n-9 5.90 52.53

FEED A & LPC FEED B & LPC POM & LPC OM & LPC FA Contrib. Cum. FA Contrib. Cum. FA Contrib. Cum. FA Contrib. Cum. % % % % % % % %

18:1n-9 14.48 14.48 16:1n-9 11.87 11.87 20:5n-3 15.74 15.74 20:5n-3 13.72 13.72 22:6n-3 10.01 24.50 14:0 10.21 2208 18:1n-11 8.25 23.99 20:1n-9 8.39 22.10 16:1n-9 9.86 34.36 20:2n-6 10.18 32.27 14:0 7.91 31.90 18:1n-7 6.77 28.87 18:2n-6 9.62 43.98 16:0 7.99 40.25 18:2n-6 7.76 39.66 16:0 6.32 35.19 14:0 7.95 51.93 18:2n-6 6.00 46.25 18:1n-7 6.95 46.62 22:4n-6 6.31 41.50 18:1n-11 5.90 52.15 16:1n-7 5.71 52.52 22:5n-3 6.23 47.74 22:1n-9 5.89 53.63

larities recorded between the FA profiles of SPC, larity between LPC and SPC); and 20:4n-6 (explain- MPC and LPC in sand bed tanks were as follows: ing 10.8% of the dissimilarity between MPC and 6.1% for LPC vs. MPC; 8.2% for LPC vs. SPC; and SPC). 5.7% for MPC vs. SPC. The single FAs contributing Regarding the FA profiles of LPC and potential the most to the re corded dissimilarities were: 16:0 sources of food (POM, OM, FEED A and B), SIMPER (explaining 8.9% of the dissimilarity between LPC revealed a dissimilarity of 40.9% between LPC and and MPC); 18:2n-6 (explaining 9.5% of the dissimi- POM, 60.8% between LPC and OM, and 43.7 and 44.3% between LPC and FEED A and B, respectively (Table 2). A first hierarchical cluster analysis (Fig. 1) re vealed 2 distinct clusters (with a 90% similarity) separating the FA profiles of WP Hediste diversicolor from those of conspecifics cultured in sand bed tanks. A second hierarchical cluster analysis (Fig. 2) showed that the FA profiles displayed by OM sam- ples were clearly separated (similarity <60%) from those exhibited by cul- tured, large H. diversicolor and POM, as well as FEED A and B. The CAP analysis revealed that the FA profile most closely resembling that of LPC was that of POM, as 100% of all LPC profiles were allocated to Fig. 1. Hierarchical cluster analysis groups of the fatty acid profiles of wild POM when LPC was selected for the (WP), and small, medium and large cultured Hediste diversicolor (SPC, MPC ‘leave-one-out’ allocation routine and LPC, respectively) (Table 3). Marques et al.: Ragworm cultivation and added value 85

The FA profiles of cultivated specimens (SPC, MPC and LPC) and wild polychaetes (WP) are in agreement with previous studies (Table 4) reporting that the most abundant FA recorded in cul- tured and wild polychaetes are pal mitic acid, stearic acid, oleic acid and EPA (García-Alonso et al. 2008, Bischoff et al. 2009, Techaprempreecha et al. 2011, Lillebø et al. 2012). Of these, EPA (8.34 ± 0.36 µg g−1 DW in LPC and 5.51 ± 0.29 µg g−1 DW in WP) was the most abundant HUFA present in polychaete biomass. It can be seen that cultivated speci- Fig. 2. Hierarchical cluster analysis groups of the fatty acid profiles of large Hediste mens are able to incorporate al- diversicolor (LPC) cultured in the sand bed tanks, organic matter (OM), particulate most all available EPA in their organic matter (POM) and fish feed (FEED A and FEED B) food sources into their tissues, even when this FA is present in DISCUSSION low levels (e.g. 0.11 ± 0.04 µg g−1 DW for POM, 0.36 ± 0.12 µg g−1 DW for OM, 7.06 ± 0.71 µg g−1 DW for The reproduction success of wild Hediste diversi- FEED A and 16.20 ± 2.15 µg g−1 DW for FEED B). EPA color under cultivated conditions (i.e. using RAS ef - levels present in food sources can be complemented fluents displaying a high level of uneaten/ undigested by the polychaetes through de novo synthesis (Santos feed and fish feces) confirms the ability of this species et al. 2016). This FA is one of the major components of to switch its feeding behavior according to trophic/ fish oil, a precursor of prostaglandins and thrombox- environmental conditions (Bischoff et al. 2009). In ane. It cannot be synthesized de novo in humans fact, it has been demonstrated that under nutrient- (García-Alonso et al. 2008), and is therefore classified enrichment conditions, the surface deposit-feeding as an essential FA (Olive 1999). DHA is also an essen- behavior of H. diversicolor is enhanced over suspen- tial FA of great importance in marine fish nutrition. sion feeding and/or predation (Aberson et al. 2016). The significant increase in DHA concentrations in The increment recorded in biomass and density dur- cultured specimens may be due to a selective reten- ing the experimental period corroborates that the tion in tissues and/or their ability to elongate FAs available food was sufficient to secure growth and through a pathway that involves chain elongation of reproduction. In addition, the bentho-pelagic life EPA and its desaturation to obtain DHA (Olive et al. cycle of H. diversicolor and its direct development in 2009). The absence of DHA in the tissues of wild the sand bed tank environment, without loss of off- specimens might reflect a deficiency of this FA in nat- spring, allows reproductive success and a significant ural intertidal mud-flat food sources. Arachidonic increase in polychaete density. acid (AA, 20:4n-6) is another n-6 long-chain FA es-

Table 3. Cross validation success of large Hediste diversicolor (LPC) fatty acid (FA) profile based on FA profiles of organic matter (OM), particulate organic matter (POM) and fish feed (FEED A and FEED B)

Allocation of observations to groups Total per group % correct OM POM FEED A FEED B

OM 6 0 0 0 6 100 POM 0 6 0 0 6 100 FEED A 0 0 3 0 3 100 FEED B 0 0 0 3 3 100 LPC 0 6 0 0 6 100 86 Aquacult Environ Interact 10: 79–88, 2018

sential to fish diets. AA was detected in all samples of polychaetes (WP, SPC, MPC and LPC with 0.84 ± 0.10, 0.99 ± 0.33, 1.25 ± 0.15 and 1.61 ± 0.17 µg g−1 DW, respectively) and takes part in several metabolic pathways in invertebrates, e.g. Perinereis nuntia (Techaprempreecha et al. 2011) and Nereis virens (Brown et al. 2011). This FA can also be biosynthe- sized from linoleic acid (Bischoff et al. 2009). It has -7 -3 -3 Santos et al. (2016) -3 been shown that, to some extent, polychaetes (Areni- n n n n -9 Bischoff et al. (2009) -9 -9 -9 -7 Lillebø et al. (2012) Reference n n n n n cola marina) have the ability to elongate 18:2n-6 5 Techaprempreecha (linoleic acid) to produce 20:2n-6, which through a desaturation pathway can yield de novo 20:4n-6 -6 18:0 (2008) -6 C22:6 -9 16:1 Brown et al. (2011) -3 20:1 -9 20:1 -9 20:1 -6 18:1 -6 C18:1 n n n n n n n n (Olive et al. 2009). In addition, polychaetes can also retain most 20:4n-6 (AA) present in their diet (OM = 0.12 ± 0.02 µg g−1 DW). AA is an essential FA in fish -12 18:0 20:1

n nutrition, particularly during early life phases (e.g. -6 18:0 C22:6 -9/ -9 18:2 -9 20:1 17:0 -9 18:0 15:0 García-Alonso et al. -7 16:1 18:0 -9 18:2 -9 18:0 C22:6 -3 18:1 -9 20:5 -3 18:1 11 18:1 larval stages), thus being paramount when selecting n n n n n n n n n n n n ingredients to fulfil the nutritional needs of cultured marine fish (Bell & Sargent 2003). According to other studies, palmitic acid (16:0) represents one of the -3 18:1 -3 18:1 -3 18:1 -3 18:1 -6 20:5 -11 18:1 -11 20:5 -3 22:1 n n n n n n n n most abundant saturated FAs present in polychaetes (García-Alonso et al. 2008, Brown et al. 2011, Techa- prempreecha et al. 2011, Santos et al. 2016). Palmitic -3 16:0 20:4 -3 16:0 18:1 -3 16:0 18:1 -3 16:0 18:1 -3 16:0 18:0 20:4 -3 16:0 18:0 20:4 20:5 n n n n n n acid is the first FA to be biosynthesized and is a pre- 6:0 18:2 18:1 18:0 20:2 et al. (2011) 0:5 cursor of longer-chain saturated FA (Nelson & Cox 2004). Palmitic acid is also a precursor of many types 16:0 22:1 16:0 22:1 16:0 20:5 16:0 20:5 of mole cules with physiological relevance, such as membrane lipids, fats and waxes (García-Alonso et al. 2008). Results show that the proposed system for culturing H. diversicolor enables the recovery of HUFAs (e.g.

(Laranjo basin) 20:5 EPA, DHA and AA) and palmitic acid into the tissues

(Laranjo basin) 20:5 of these polychaetes. These FA can be reintroduced into productive systems through the potential of H. diversicolor to recycle these key ingredients avail- able in different food sources originating from culti- vated fish (POM, OM, FEED A and B). Without the Bolboschoenus Juncus maritimus action of H. diversicolor, these essential FA would likely be lost into the environment. In its natural habitat, H. diversicolor is prey for ) higher trophic levels, either small fish such as the common goby Pomatoschistus microps and the sand goby P. minutus (Scaps 2002), or larger fish also used

Table 4. SummaryTable table with the most abundant fatty acids in polychaete worms with different food sources for human consumption, such as the Senegalese sole

Sparus aurata Solea senegalensis (Rosa et al. 2008). This may some- Commercial diet (seabream dry feed) 16:0 20:5 Commercial worm diet 16:0 18:2 Wild (Bangphra beach. Chonburi providence,Wild Thailand) 16:0 18:1 18:0 16:1 20: Sedimentation tank (OM, feces and uneaten food of 20:5 Mud flats (Cardiff Bay) 20:5 Mud flats colonized by how explain the high level of acceptability displayed by this prey when offered to farmed fish species. In the present study, it was shown that cultivated H. diversicolor displays a FA profile that holds great potential to marine fish aquaculture as: (1) it is able to reduce the loss of essential FA from the productive (O.F. Müller, 1776) Müller, (O.F. Fish food 16:0 1776) Müller, (O.F. (O.F. Müller, 1776) Müller, (O.F. Commercial diet (semi-wet pellets for cultured sole) 16:0 20:5 (O.F. Müller, 1776) Müller, (O.F. Mud flats colonized by system to the environment; and (2) it displays an Eel sludge Without vegetation (Laranjo basin) Without 20:5 Fecal waste Commercial shrimp diet 1 Wild (French Wild Atlantic coast) 2 Mixed waste Pellet waste Hediste diversicolor Nereis virens Hediste diversicolor Perinereis nuntia Hediste diversicolor Polychaete worm Food source 5 of the most abundant fatty acids Top Hediste diversicolor enhanced nutritional value when compared to wild Marques et al.: Ragworm cultivation and added value 87

specimens (e.g. the presence of DHA in its FA pro- Aberson MJ, Bolam SG, Hughes RG (2016) The effect of file). Moreover, the average number of LPC recorded sewage pollution on the feeding behaviour and diet of Hediste (Nereis diversicolor (O.F. Müller, 1776)) in three in the present study after 5 mo is the equivalent of estuaries in south-east England, with implications for approximately 100 bait packages identical to those saltmarsh erosion. Mar Pollut Bull 105: 150−160 initially used to stock each tank (with circa 70 speci- Alexander KA, Potts TP, Freeman S, Israel D and others mens per package). (2015) The implications of aquaculture policy and regu- lation for the development of integrated multi-trophic aquaculture in Europe. Aquaculture 443:16−23 Anderson MJ (2008) Animal-sediment relationships re-vis- CONCLUSIONS ited: characterising species’ distributions along an envi- ronmental gradient using canonical analysis and quan- Overall, the present study confirms the potential of tile regression splines. J Exp Mar Biol Ecol 366: 16−27 Aued-Pimentel S, Lago JHG, Chaves MH, Kumagai EE Hediste diversicolor for the bioremediation of super- (2004) Evaluation of a methylation procedure to deter- intensive marine fish farm effluents and highlights its mine cyclopropenoid fatty acids from Sterculia striata ability to retain high value nutrients (e.g. HUFA in St. Hil. Et Nauds seed oil. J Chromatogr A 1054:235−239 general and EPA in particular, and to a lesser extent Barrington K, Chopin T, Robinson S (2009) Integrated multi- trophic aquaculture (IMTA) in marine temperate waters. DHA) from fish feeds that would otherwise be lost In: Soto D (ed) Integrated mariculture: a global review. from the production environment. Ragworm biomass FAO Fish Aquacult Tech Pap 529. FAO, Rome, p 7−46 of large specimens may be valued selectively if Bell JG, Sargent JR (2003) Arachidonic acid in aquaculture traded live for sport fish bait, as these are traded at a feeds: current status and future opportunities. Aquacul- unitary level (pack of live bait) and not per kg. LPC ture 218: 491−499 Bischoff AA, Fink P, Waller U (2009) The fatty acid composi- can also be frozen and traded for maturation diets for tion of Nereis diversicolor cultured in an integrated recir- fish and/or shrimp broodstock, with their higher level culated system: possible implications for aquaculture. of DHA being used as a feature that may grant them Aquaculture 296: 271−276 a premium market value over wild conspecifics. The Breton S, Dufresne F, Desrosiers G, Blier PU (2003) Popula- tion structure of two northern hemisphere polychaetes, biomass of small, medium or large cultured poly- Neanthes virens and Hediste diversicolor (Nereididae), chaetes (SPC, MPC and LPC, respectively) may be with different life-history traits. Mar Biol 142: 707−715 valued as a whole through their use as a premium Brown N, Eddy S, Plaud S (2011) Utilization of waste from a ingredient for finishing diets of farmed marine marine recirculating fish culture system as a feed source for the polychaete worm, Nereis virens. Aquaculture organisms, thus overcoming the need to grade cul- 322−323: 177−183 tured ragworms. Chopin T, Robinson SMC, Troell M, Neori A, Buschmann AH, Fang J (2008) Multitrophic integration for sustain- able marine aquaculture. In: Jørgensen SE, Fath BD Acknowledgements. This research was supported by the (eds) Ecological engineering. Encyclopedia of ecology, PROMAR–European Fisheries Fund through the STEP pro- Vol. 3. Elsevier, Oxford, p 2463−2475 ject, coordinated by Aquacria Piscícolas SA (Sea8 Company) Clarke K, Gorley R (2006) PRIMER v6: User manual/tutorial. and by FCT (Portuguese Foundation for Science and Tech- PRIMER-E, Plymouth nology) through a PhD grant awarded to B.M. (SFRH/BD/ Fang J, Jiang Z, Jansen H, Hu F and others (2017) Applica- 96037/2013) and a postdoctoral grant awarded to C.N. bility of Perinereis aibuhitensis Grube for fish waste (SFRH/ BPD/100627/2014). This work was also supported by removal from fish cages in Sanggou Bay, PR China. the Integrated Programme of SR&TD ‘Smart Valorization of J Ocean Univ China 16:294−304 Endogenous Marine Biological Resources Under a Chang- Fidalgo e Costa P, Narciso L, da Fonseca LC (2000) Growth, ing Climate’ (Centro-01-0145-FEDER-000018), co-funded survival and fatty acid profile of Nereis diversicolor (O.F. by the Centro 2020 program, Portugal 2020, European Müller, 1776) fed on six different diets. Bull Mar Sci 67: Union, through the European Regional Development Fund. 337−344 Thanks are also due, for financial support, to CESAM Fritzsche D, von Oertzen JA (1995) Metabolic responses to (UID/AMB/50017/2013) and QOPNA (FCT UID/QUI/00062/ changing environmental conditions in the brackish 2013), to FCT/MEC through national funds, and to FEDER water polychaetes Marenzelleria viridis and Hediste for the co-funding within the PT2020 Partnership Agree- diversicolor. Mar Biol 121: 693−699 ment and Compete 2020. The authors also thank all col- García-Alonso J, Müller CT, Hardege JD (2008) Influence of leagues who helped with the field work. food regimes and seasonality on fatty acid composition in the ragworm. Aquat Biol 4: 7−13 Hutchinson W, Jeffrey M, O’Sullivan D, Casement D, Clarke LITERATURE CITED S (2004) Recirculating aquaculture systems minimum standards for design, construction and management. Aberson MJR, Bolam SG, Hughes RG (2011) The dispersal RAS MSDCM V2 040220. Inland Aquaculture Associa- and colonisation behaviour of the marine polychaete tion of South Australia, Kent Town Nereis diversicolor (O. F. Müller) in south-east England. Lillebø AI, Cleary DFR, Marques B, Reis A, Silva TL, Calado Hydrobiologia 672:3−14 R (2012) Ragworm fatty acid profiles reveals habitat and 88 Aquacult Environ Interact 10: 79–88, 2018

trophic interactions with halophytes and with mercury. Mar Pollut Bull 64: 2528−2534

Luis OJ, Passos AM (1995) Seasonal changes in lipid content 0.14 and composition of the polychaete Nereis (Hediste) diversicolor. Comp Biochem Physiol 111: 579−586 Marques B, Calado R, Lillebø AI (2017) New species for the biomitigation of a super-intensive marine fish farm efflu- ent: combined use of polychaete-assisted sand filters and halophyte aquaponics. Sci Total Environ 599−600: 1922−1928 Martins CIM, Eding EH, Verdegem MCJ, Heinsbroek LTN

and others (2010) New developments in recirculating Hediste diversicolor aquaculture systems in Europe: a perspective on envi- ronmental sustainability. Aquacult Eng 43: 83−93 Murray F, Solan M, Douglas A (2017) Effects of algal enrich- ment and salinity on sediment particle reworking activity and associated nutrient generation mediated by the intertidal polychaete Hediste diversicolor. J Exp Mar Biol Ecol 495: 75−82 Nelson DL, Cox MM (2004) Lehninger principles of bio- chemistry, 4th edn. W H Freeman & Co, New York, NY 20.4 ± 0.1 20.4 ± 0.1 18.5 ± 0.1 18.9 ± 0.2 Nesto N, Simonini R, Prevedelli D, Da Ros L (2012) Effects of N–C SBT C SBT N–C SBT C SBT diet and density on growth, survival and gametogenesis of Hediste diversicolor (O.F. Müller, 1776) (Nereididae, Polychaeta). Aquaculture 362−363:1−9 ± 0.0 21.2 ± 0.0 21.2 ± 0.0 21.1 ± 0.0 21.1 ± 0.0 Olive PJW (1999) Polychaete aquaculture and polychaete science: a mutual synergism. Hydrobiologia 402:175−183 Olive PJW, Duangchinda T, Ashforth E, Craig S, Ward AC, Davies SJ (2009) Net gain of long-chain polyunsaturated fatty acids (PUFA) in a lugworm Arenicola marina biotur- ± 0.18 7.90 ± 0.08 7.71 ± 0.08 7.56 ± 0.03 7.93 ± 0.04 7.95 ± 0.03 bated mesocosm. Mar Ecol Prog Ser 387:223−239 Palmer PJ (2010) Polychaete-assisted sand filter. Aquacul- ture 306: 369−377 Rosa S, Granadeiro JP, Vinagre C, França S, Cabral HN, Palmeirim JM (2008) Impact of predation on the poly- APPENDIX chaete Hediste diversicolor in estuarine intertidal flats. Estuar Coast Shelf Sci 78: 655−664 Santos A, Granada L, Baptista T, Anjos C and others (2016) Effect of three diets on the growth and fatty acid profile of the common ragworm Hediste diversicolor (O.F. Müller, 1776). Aquaculture 465: 37−42 Scaps P (2002) A review of the biology, ecology and poten- tial use of the common ragworm Hediste diversicolor (O. F. Müller) (Annelida: Polychaeta). Hydrobiologia 470: 203−218 Schneider O, Sereti V, Eding EH, Verreth JAJ (2005) Analy- sis of nutrient flows in integrated intensive aquaculture systems. Aquacult Eng 32: 379−4011 Smith RI (1964) On the early development of Nereis diversi- color indifferent salinities. J Morphol 114:437−463 Techaprempreecha S, Khongchareonporn N, Chaicharoen- pong C, Aranyakananda P, Chunhabundit S, Petsom A (2011) Nutritional composition of farmed and wild sand- worms, Perinereis nuntia. Anim Feed Sci Technol 169: ) 9.69 ± 0.34 8.90 ± 0.31 8.80 ± 0.04 7.96 ± 0.15 8.40 ± 0.49 7.77 ± 0.25 8.45 ± 0.30 7.08 ± 0.04 9.00 ± 0.31 9.09 ±

265−269 −1 van Rijn J, Tal Y, Schreier HJ (2006) Denitrification in recir- culating systems: Theory and applications. Aquacult Eng 34: 364−376 Würzberg L, Peters J, Schüller M, Brandt A (2011) Diet insights of deep-sea polychaetes derived from fatty acid analyses. Deep Sea Res Part II 58:153−162 Table A1. Physicochemical parameters (means ± SD) of non-colonized (N-C) and colonized (C) sand bed tanks (SBT) containing Table N–C SBT C SBT N–C SBT C SBT N–C SBT C SBT Temperature (°C)Temperature 20.5 ± 0.6 20.6 ± 0.6 18.8 ± 0.1 18.8 ± 0.0 19.4 ± 0.1Salinity 19.4 ± 0.2 21.7 ± 0.0 21.7 ± 0.0 21.0 ± 0.0 21.0 ± 0.0pH 21.2 ± 0.0 21.2 – 8.95 ± 0.12 –Dissolved oxygen (mg l – 8.04 Parameter Month 1 Month 2 Month 3 Month 4 Month 5

Editorial responsibility: Pablo Sánchez Jerez, Submitted: September 22, 2017; Accepted: December 11, 2017 Alicante, Spain Proofs received from author(s): February 1, 2018